JP5737118B2 - Storage system and storage system network - Google Patents

Description

  The present invention relates to a storage system.

  FIG. 1 shows the configuration and usage of an existing storage system 200 having a remote copy function. Note that the storage system 200 described below is a system (apparatus) that can increase the total storage capacity by increasing the number of storage devices 220 (usually hard disk drives) mounted on the chassis. The storage system 200 is also a system in which a specific storage device 220 is controlled by each of a plurality of control modules 210 mounted on the housing.

As shown in FIG. 1, the storage system 200 is used in a state of being connected to a host 300 and another storage system 200 (hereinafter referred to as a copy destination system 200) installed at a remote location. . A plurality of sets of control data storage areas and update data storage areas are provided in each control module 210 of the storage system 200 connected to the host 300 (hereinafter referred to as the copy source system 200). Also, a plurality of sets of control data storage areas and update data storage areas are provided in each control module 210 of the copy destination system 200. The control module 210 is a unit combining a CPU (Central Processing Unit), a RAM (Random Access Memory), etc.
Both the control data storage area and the update data storage area are one storage area on the RAM.

  A set of control data storage areas and update data storage areas in each control module 210 are associated with each other as storage areas belonging to one buffer set. Each control module 210 in the copy source system 200 operates as follows.

  When each control module 210 receives a write request for certain update data from the host 300, each control module 210 stores the write destination information and stores it in a control data storage area belonging to a certain buffer set (hereinafter referred to as buffer set 0). After storing the control data, it responds to the write request. Here, the write destination information is information indicating a write position of update data. The control data is information indicating a write position in the copy destination system 200 for update data.

  Also, each control module 210 in the copy source system 200 performs a process of reading update data from the storage device 220 based on the write destination information and storing it in the update data storage area belonging to the buffer set 0 as needed (as appropriate).

  Furthermore, each control module 210 in the copy source system 200 switches the buffer set synchronously when new control data cannot be stored in the control data storage area in any of the control modules 210. That is, each control module 210 is in a state of storing control data and update data in, for example, a control data storage area and an update data storage area belonging to the buffer set 1, respectively.

  Then, after the storage of the update data in the update data storage area belonging to the buffer set 0 is completed, each control module 210 in the copy source system 200 sequentially stores information on each storage area belonging to the buffer set 0. It is transmitted to the copy destination system 200.

  Each control module 210 in the copy destination system 200 temporarily stores the received information as being related to the buffer set 0. Each control module 210 in the copy destination system 200 receives information (a plurality of update data and control data related to each update data) when all information related to the buffer set 0 is prepared in the own system 200. ) Is written to the storage device 220.

JP 2006-260292 A International Publication No. 2008-152691 Pamphlet JP 2010-039574 A JP 2011-086160 A JP 2011-086006 A Japanese Patent Laid-Open No. 2003-167683 Japanese Patent Laid-Open No. 7-244597 JP 2003-122509 A

  In short, if the above-described storage system 200 (hereinafter referred to as the existing SS 200) is used in the form shown in FIG. 1, data in the copy destination system 200 is periodically backed up in the copy source system 200. The environment can be realized. Therefore, if the existing SS 200 is used in the form shown in FIG. 1, when the copy source system 200 becomes unusable due to a disaster or the like, the business can be resumed using the copy destination system 200.

  However, in the existing SS 200, when an amount of storage capacity that cannot be accommodated by addition to the storage device 220 is required, the system itself must be replaced with a larger capacity.

  Specifically, if another existing SS 200 is connected to the host 300, the storage capacity on the host 300 side (the storage capacity that the host 300 can use for storing data) can be expanded. If another existing SS 200 (hereinafter referred to as the second copy destination system 200) is added to the remote location, the existing SS 200 (hereinafter referred to as the second copy source system 200) added to the host 300 side will be added. Can be backed up in the second copy destination system 200.

  However, the remote copy between the two added systems is completely independent of the remote copy between the original copy source / copy destination system 200 (hereinafter referred to as the first copy source / first copy destination system 200). Is called. Accordingly, in a system in which four existing SSs 200 are combined, the data in the first copy destination system 200 is backup data in the first copy source system 200 at a certain time T1, and the second copy destination system 200. May be backup data in the second copy source system 200 at another time T2 (for example, a time different from T1 by 30 minutes).

If the backup data in the first and second copy destination systems 200 is such, the business cannot be resumed using the first and second copy destination systems 200 (for example, the first and second copy There is a case where the database is distributed and stored in the original system 200). For this reason, when the existing SS 200 is used, when the amount of storage capacity that cannot be accommodated by addition to the storage device 220 is required, the system itself must be replaced with a larger capacity. It is.

  Therefore, an object of one aspect of the disclosed technology is to provide a storage system capable of remote copy in cooperation with another storage system connected to the same host.

  Another object of the present invention is to provide a storage system network in which a plurality of all host-side storage systems function as a single storage system and remote copy is performed to a plurality of backup storage systems. To do.

  In order to solve the above problem, a storage system connected to a host and a backup storage system according to an aspect of the disclosed technology includes a first storage unit, and a second storage unit having a plurality of update data storage areas. A control unit that receives an update data write request to the first storage unit from the host, and the control unit is a processing target storage area that is one of the plurality of update data storage areas when the write request is received In addition, it is determined whether or not there is a free area having a capacity capable of storing the update data, and when it is determined that there is a free area, the update data write area in the first storage unit is indicated. 1 write destination information, second write destination information indicating update data write area in the backup storage system, and update data storage area used as processing target storage area Control information including identification information is stored in the second storage unit, and update data is written to the first storage unit. When it is determined that there is no free area, the processing target storage area is stored in the second storage unit. Change to another update data storage area, store the control information in the second storage unit, write the update data in the first storage unit, and change the processing target storage area to another storage system connected to the host. Asynchronously with the processing of the write request, in accordance with the control information stored in the second storage unit, the update data already written in the first storage unit is read into each update data storage area of the second storage unit and read out. The update data on the completed update data storage area and the second write destination information related to the update data are transmitted to the backup storage system.

Further, according to another aspect of the disclosed technology, a storage system network including a plurality of host-side storage systems connected to a host and a plurality of backup storage systems, wherein the storage systems are communicably connected. Each of the host-side storage systems includes a first storage unit, a second storage unit having a plurality of update data storage areas, and a first control unit that receives an update data write request to the first storage unit from the host. The first control unit has a free area with a capacity capable of storing update data in the processing target storage area that is one of the plurality of update data storage areas when the update data write request is received. First write destination information indicating an update data write area in the first storage unit when it is determined whether there is a free area Control information including second write destination information indicating the update data write area in the backup storage system and identification information of the update data storage area used as the processing target storage area, the second storage unit And when the update data is written to the first storage unit and it is determined that there is no free area, the processing target storage area is changed to another update data storage area in the second storage unit, and the control information Is stored in the second storage unit, the update data is written in the first storage unit, the other storage systems on the other host side are instructed to change the processing target storage area, and the second storage is performed asynchronously with the write request processing. In accordance with the control information stored in the unit, the update data written in the first storage unit is read into each update data storage area of the second storage unit, and the update data that has been read out And a second write destination information about the update data and the update data on 憶領 zone includes a system to the storage system for backup. Further, the storage system network receives the update data and the second write destination information transmitted by the third storage unit and the host storage system associated with the backup storage system as each backup storage system, And a second control unit that writes the received update data to a storage area in the third storage unit indicated by the received second write destination information.

  The disclosed technology includes a storage system capable of remote copying in cooperation with other storage systems connected to the same host, a plurality of host-side storage systems, and a plurality of backup storage systems, It is possible to provide a storage system network in which remote copying from a host-side storage system to a corresponding backup storage system is performed cooperatively.

FIG. 1 is an explanatory diagram of the configuration and functions of an existing storage system. FIG. 2 is an explanatory diagram of a configuration of a storage system according to an embodiment of the disclosed technology and a configuration of a storage system network constructed using the storage system according to the embodiment. FIG. 3 is an explanatory diagram of a group management table prepared on the memory of each CM in the storage system according to the embodiment. FIG. 4A is a flowchart (part 1) of the initial construction negotiation process executed in the storage system network according to the embodiment. FIG. 4B is a flowchart (part 2) of the initial construction negotiation process executed in the storage system network according to the embodiment. FIG. 5 is an explanatory diagram of a buffer set management table prepared on the memory of the reception CM. FIG. 6 is an explanatory diagram of an empty reception buffer table prepared on the memory of the main reception master CM. FIG. 7 is an explanatory diagram of a buffer set management table prepared on the memory of the transmission CM. FIG. 8 is an explanatory diagram of the buffer management table. FIG. 9 is an explanatory diagram of an initial copy session management table. FIG. 10 is a flowchart of processing executed in the storage system network according to the embodiment at the time of copy range setting work. FIG. 11 is a flowchart of the initial copy process. FIG. 12 is a flowchart of a write request analysis / response process executed when a write request is received. FIG. 13 is a flowchart of the storage reservation / queuing process. FIG. 14 is an explanatory diagram of a write request copy control table. FIG. 15 is a flowchart of data storage processing. FIG. 16 is a flowchart of buffer set switching request processing. FIG. 17A is a flowchart (part 1) of the buffer set switching process. FIG. 17B is a flowchart (part 2) of the buffer set switching process. FIG. 18 is a flowchart of processing executed by the CM in charge of the copy session that has received the switch completion notification. FIG. 19 is a flowchart of the empty reception buffer ID assignment process. FIG. 20A is a flowchart (part 1) of the buffer set management table transfer process. FIG. 20B is a flowchart (part 2) of the buffer set management table transfer process. FIG. 20C is a flowchart (part 3) of the buffer set management table transfer process. FIG. 20D is a flowchart of processing executed when a buffer set reception completion notification is received. FIG. 21A is a flowchart (part 1) of the buffer management table and buffer transfer processing. FIG. 21B is a flowchart (part 2) of the buffer management table and buffer transfer processing. FIG. 21C is a flowchart (part 3) of the buffer management table and buffer transfer processing. FIG. 21D is a flowchart (part 4) of the buffer management table and buffer transfer processing. FIG. 21E is a flowchart (part 5) of the buffer management table and buffer transfer processing. FIG. 21F is a flowchart (part 6) of the buffer management table and buffer transfer processing. FIG. 21G is a flowchart (part 7) of the buffer management table and buffer transfer processing. FIG. 22A is a flowchart (part 1) of the expansion process. FIG. 22B is a flowchart (part 2) of the expansion process. FIG. 22C is a flowchart (part 3) of the expansion process. FIG. 23A is a flowchart (part 1) of the release process. FIG. 23B is a flowchart (part 2) of the release process. FIG. 24 is a flowchart of buffer set regular switching processing.

  Hereinafter, embodiments of the technology disclosed in this specification will be described in detail with reference to the drawings.

  FIG. 2 shows a configuration of the storage system 100 according to an embodiment of the disclosed technology and a configuration of a storage system network constructed by combining a plurality of storage systems 100.

  First, an overview of a storage system (hereinafter also referred to as SS) 100 according to the present embodiment and an overview of an SS network constructed by combining a plurality of SSs 100 will be described with reference to FIG.

As shown in FIG. 2, the SS 100 according to the present embodiment is an apparatus (system) including a housing 105, one or more CMs (Controller Modules) 110, and a storage unit 120.

  The casing 105 can be equipped with auxiliary storage devices (hard disk devices, SSDs (Solid State Drives), etc.) 125 of various specifications (storage capacity, response speed, etc.) up to the first specified number, and CM 110 can be mounted up to the second specified number. Is a unit.

  The storage unit 120 is a unit realized by a plurality of auxiliary storage devices 125 mounted on the housing 105. At the time of SS100 setup, the user performs setting work to divide a plurality of auxiliary storage devices 125 (that is, the storage unit 120) mounted on the housing 105 into several RAID (Redundant Array of Independent Disks) groups. Is called. Also, setting work for dividing each RAID group into several logical units and work for setting (specifying) the CM 110 that is in charge of controlling each RAID group are also performed.

The CM 110 is basically a unit that performs processing (read / write access) on the contents specified by the read / write request from the host 300 on the storage unit 120. The read / write request received by each CM 110 in the SS 100 from the host 300 designates a LUN (Logical Unit Number), an LBA (Logical Block Address), and a storage area in the storage unit 120 to be read / written by the number of blocks. It is a request to do. In addition, the host 300 transmits a read / write request for a certain logical unit to the CM 110 in charge of controlling the logical unit.

As shown in FIG. 2, the CM 110 includes a CA (Channel Adapter) 113 that is an interface module for the host 300 and two FC (Fibre Channel) 115 that are interface modules for the storage unit 120. CM
110 also includes an RA (Remote Adapter) 114, which is an interface module for another SS 100, and a LAN port (not shown) for connecting the own CM 110 to a LAN (Local Area Network). Further, the CM 110 includes a ROM (Read Only Memory; not shown), a CPU (Central Processing Unit) 111 that performs various processes according to a program (firmware) stored in the ROM, and a memory 112 that is used for temporary storage of various data. It has.

  An SS network constructed using a plurality of SSs 100 includes N (≧ 2) SSs 100 (hereinafter referred to as transmission system 100) connected to the host 300 and N SSs 100 (hereinafter referred to as “transmission system 100”) installed in remote locations. And a system connected to the receiving system 100).

  As will be described in detail later, this SS network is configured such that N transmission systems 100 on the host 300 side function as one storage system, and the N receiving systems 100 on the remote side are connected to the host. This is a system in which an orderly copy of write data by 300 is performed.

  In the SS 100 according to the present embodiment, if the transmission system 100 and the reception system 100 are added to a system including several transmission systems and reception systems, it is unnecessary to replace an existing transmission / reception system. Thus, the apparatus can construct an SS network having the above functions.

  Hereinafter, the configuration and operation of the SS 100 according to the present embodiment will be described more specifically with a focus on the case where it is used as a component of the SS network.

  The SS network is normally constructed by adding several SSs 100 (receiving system 100, transmitting system 100) to an operating system including several SSs 100. Therefore, during normal construction of the SS network, the user performs setup work (setting work for a RAID group or the like) for some of the added SSs 100.

  Thereafter, the user performs configuration setting work and copy range setting work.

  First, the contents of the configuration setting work and the functions of the CM 110 related to the work will be described.

  The CM 110 for the SS 100 has a Web server function for providing various Web pages for configuration settings. Therefore, in the configuration setting work, a computer is connected to a CM 110 in a certain receiving system via a LAN (usually via a hub connected to the LAN port of each CM 110 in the receiving system), and a Web browser of the computer is started. Is done.

When a certain SS 100 is activated for the first time, each CM 110 in the SS 100 has an SS
CM-ID (numerical information in the present embodiment), which is identification information in 100, is given. Then, the CM 110 assigned with the smallest CM-ID in the SS 100 operates as the CM 110 (hereinafter referred to as a master CM) having the Web server function (and other various functions). Each of the other CMs 110 operates as a CM 110 that does not have the Web server function (hereinafter referred to as a slave CM). Accordingly, it is the master CM (CM 110 operating as the master CM) that actually receives the information set by the user.

  Although detailed explanation (configuration explanation of each web page for configuration setting etc.) is omitted, at the time of configuration setting work, identification information of each SS 100 as an SS network component (transmission system 100 or reception system 100) is set by the user. Is done. In addition, one transmission system 100 that operates as the main transmission system 100 and one or more transmission systems 100 that operate as the sub transmission system 100 are also set. One receiving system 100 that operates as the main receiving system 100 and one or more receiving systems 100 that operate as the sub receiving system 100 are also set.

  In addition, a correspondence relationship between the transmission system 100 and the reception system 100 is also set. The setting of the correspondence relationship is that a group ID is assigned to each reception / transmission system 100, and then a group ID assigned to a certain transmission system 100 and a group ID assigned to the reception system 100 associated with the transmission system 100 are set. This is done by associating.

  Further, the transmission side total buffer size, the reception side total buffer size, and the buffer set automatic switching time (details of each information will be described later) are also set by the user.

  When the configuration setting work as described above is performed, from the master CM that has actually accepted the setting, each slave CM in the own system 100 and each CM 110 in each other system 100 (transmission / reception system 100). The settings are notified. Each CM 110 in the SS network holds information for managing the setting contents by the user, the status of each buffer set, and the information for managing the execution status of various processes in its own memory 112. A group management table is prepared to keep it.

  The operation of each CM 110 in each SS 100 constituting the SS network differs depending on the type of its own system and whether it is a master CM or a slave CM. Therefore, in the following, for convenience of explanation, a master CM in the system 100 (˜ = main transmission, sub transmission, transmission, etc.) is referred to as a master CM (main transmission master CM, sub reception master CM, transmission master CM, etc. ). Similarly, the slave CM and CM 110 (master CM or slave CM) in the system 100 are denoted as ~ slave CM (main transmission slave CM or the like) and ~ CM (transmission CM or the like), respectively.

  FIG. 3 shows an example of the group management table 150 prepared by the main transmission master CM on its own memory 112 at the time of configuration setting work.

The group management table 150 shown in FIG. 3 is the main transmission master when the following settings are made by the user and the total number of CMs 110 in the two transmission systems 100 designated by the user is 4. The CM is prepared on the memory 112.
Number of transmission systems: 2
Number of receiving systems: 2
Identification information of SS 100 as the main transmission system (transmission system 0): “A”
Identification information of SS 100 as a sub transmission system (transmission system 1): “B”
Identification information of SS 100 as the main receiving system (receiving system 0): “C”
Identification information of SS 100 as a sub-reception system (reception system 1): “D”
Group ID of main transmission system: “0”
Group ID of sub transmission system: “2”
Group ID of the main receiving system: “0”
Group ID of the sub transmission system: “1”
The main transmission system is associated with the main reception system (the transmission system with the group ID “0” is associated with the reception system with the group ID “0”).
The sub transmission system is associated with the sub transmission system (the transmission system with the group ID “2” is associated with the reception system with the group ID “1”).
Transmission side total buffer size (MB): 64
Receiving side total buffer size (MB): 64
Buffer set automatic switching time (sec): 60

  The buffer size set (stored) in the group management table 150 is preliminarily set as the size (size in MB (megabinary byte) unit) of one buffer (details will be described later) prepared on the memory 112 of each CM 110. It is a defined value. The buffer management table size is also a value determined in advance as the size (size in KB (kilobinary byte)) of one buffer management table (FIG. 8: details will be described later).

  The number of buffer sets set in the group management table 150 is information that defines the number of buffer sets prepared in each transmission system 100. In the group management table 150, as the number of buffer sets, a calculation result Ns of “transmission side total buffer size / buffer size” is set.

  The buffer set is a set of buffer set management table, buffer management table and buffer on the memory 112 in the master CM of a transmission system 100, and 1 on the memory 112 in each slave CM of the same transmission system. A set of buffer management tables and buffers. Although details of each component of the buffer set will be described later, there is also a buffer set (reception side buffer set) prepared in the reception system. Therefore, if there is a possibility of misunderstanding, a buffer set prepared in each transmission system will be referred to as a transmission-side buffer set.

  Other information that can be set in the group management table 150 includes information (group ID, identification information of the transmission system 0 (= main transmission system), group ID correspondence information, etc.) indicating the setting contents by the user, and the following information: It is.

• Buffer set automatic switching time • Buffer set periodic switching processing activation status • Negotiation status • Storage buffer set number • Buffer set switching status • Buffer set switching request status • Ns buffer set usage status (“buffer set j usage status”; j = 0 to Ns-1)
-Buffer set empty waiting state-Switching acceptance status for each transmission CM in all transmission systems ("CM10 switching acceptance status", etc.)
Expanded buffer set number Receiving state of each buffer set for each receiving system (“buffer set j receiving system k receiving state”; j = 0 to Ns−1, k = 0 to the number of receiving systems−1)

  In other words, the group management table 150 is a table in which the actual configuration (the number of information that can be actually stored) varies depending on Ns, the number of reception systems 100, and the like.

  Therefore, the main transmission master CM stores the group management table 150 prepared on the memory 112 based on the setting contents by the user and Ns calculated from the transmission side total buffer size and the buffer size when the configuration setting work is performed. Determine the actual configuration.

  Then, the main transmission master CM records the setting contents by the user having the determined configuration, and the buffer size, the buffer management table size, the number of buffer sets, and the negotiation state are 8, 12, Ns, “no negotiation”, respectively. 3 is prepared (generated) on the memory 112.

  The group management table 150 prepared on the memory 112 of each of the other CMs 110 has the same configuration as the group management table 150 shown in FIG. However, the negotiation state is information used only by the transmission master CM (the main transmission master CM and the sub transmission master CM). Further, the buffer set automatic switching time and the buffer set automatic switching processing activation state are information used only by the main transmission master CM.

  Therefore, the sub-transmission master CM prepares the group management table 150 in the memory 112 as a negotiated state in which no significant information is set. Further, each transmission slave CM and each reception CM prepare (generate) a group management table 150 in which no significant information is set as a negotiation state, a buffer set automatic switching time, and a buffer set automatic switching processing activation state. )

  Next, the operation of each CM 110 after the generation of the group management table 150 is completed will be described. In the following description, the buffer set is also expressed as BS.

  When the generation of the group management table 150 on the memory 112 is completed, each transmission master CM (master CM in each transmission system 100) periodically executes the initial construction negotiation process of the procedure shown in FIGS. 4A and 4B. It becomes a state. This initial construction negotiation process is not executed only by the transmission master CM, but is executed by the master CM and slave CM in a certain transmission system and the master CM in the reception system corresponding to the transmission system. Is. Each table that each CM 110 refers to / updates during the initial construction negotiation process and each process described below is a table on its own memory 112.

  As shown in FIG. 4A, the transmission master CM that has started the initial construction negotiation process first determines whether or not the negotiation state on the group management table 150 (FIG. 3) is “negotiated” (step S3). S1001). If the negotiation state is not “negotiated” (step S1001; NO), the transmission master CM ends this initial construction negotiation process without performing anything.

  On the other hand, when the negotiation state is “negotiation not performed” (S1001; YES), the transmission master CM changes the negotiation state on the group management table 150 to “being negotiated” (step S1002). Next, the transmission master CM transmits an initial notification having a predetermined content to the corresponding reception system 100 (the reception system 100 associated with the own transmission system 100) (step S1003).

  This initial notification is processed by a master CM (reception master CM) in the corresponding reception system 100. The receiving master CM that has received the initial notification initializes the expanded BS number on the group management table 150 to “invalid” (step S1004). In step S1004, the reception master CM initializes the reception state (“buffer set 0 reception system 0 reception state”, etc.) for each BS on the group management table 150 to “not received”. The process to do is also performed.

  Thereafter, the reception master CM stores a BS management table in which initial values are set as various information for each reception-side buffer set n (n = 0 to Ns−1: Ns is the number of BSs) on its own memory 112. Generate (step S1005).

  Here, the reception-side buffer set n is a reception-side buffer set identified by a buffer set number n. The receiving side buffer set is a set of BS management table, buffer management table and buffer on the memory 112 of the master CM in a certain receiving system, and 1 on the memory 112 of each slave CM in the same receiving system. A set of buffer management tables and a set of buffers. Ns reception side buffer sets are prepared in the reception system with BS numbers of 0 to Ns−1. The number Ns is a number when a buffer is not allocated and is counted as a receiving buffer set.

  Hereinafter, the BS management table for the reception side buffer set and the BS management table for the reception side buffer set n are also referred to as a reception side BS management table and a reception side BS management table n, respectively.

  FIG. 5 shows an example of the receiving side BS management table. As is apparent from FIG. 5, the receiving-side BS management table includes a plurality of sets of transmission buffer IDs, reception buffer IDs, buffer management table reception states, and buffer reception states, as well as BS numbers, buffer ID reception states, and deployment wait states. It is a storable table.

  In step S1005, the reception master CM prepares on the memory 112 the following information set as each reception-side BS management table n (n = 0 to “Ns−1”).

Buffer set number: n
Buffer ID status: “Not received”
Deployment waiting state: “Not waiting”
Each transmission buffer ID: “Invalid”
Each receive buffer ID: “Invalid”
Receiving status of each buffer management table: “Not received”
Each buffer reception status: “Not received”

  In step S1006, the reception master CM that has finished the processing of step 1005 first prepares it on the memory 112 of each CM 110 in the own system from the reception-side total buffer size and buffer size on the group management table 150. The number Nr of buffers is calculated. More specifically, the reception master CM calculates the “reception side total buffer size / 8” (8 is the buffer size), thereby calculating the number of buffers prepared on the memory 112 of each MC 110 in the own system 100. Nr is calculated.

  Next, the reception master CM generates buffer IDs for the Nr buffers prepared on its own memory 112 and the Nr buffers prepared on the memory 112 of each slave CM.

  When each CM 110 prepares Y buffers on the memory 112, the storage area on the memory 112 used as each buffer is ““ predetermined address + b · 8 MB ”(b = 0 to an integer value of Y−1). Is defined as a continuous storage area of 8 MB with the first address as “.” In addition, the buffer ID of “buffer with“ default address + b · 8 MB ”as the head address” on the memory 112 of a certain CM 110 is the CM-ID of the CM 110 (1 byte data in this embodiment), and b is the buffer ID. It is defined as information that continues 1 byte data to be represented.

  Accordingly, the reception master CM uses Nr data obtained by continuing 1-byte data indicating 0 to Nr-1 in each CM-ID as buffer IDs of Nr buffers prepared on its own memory 112. Is generated. Further, the reception master CM continues 1-byte data indicating 0 to Nr-1 to the CM-ID of each slave CM as the buffer ID of Nr buffers prepared on the memory 112 of each slave CM. Nr pieces of data are generated.

  As described above, the reception master CM that has generated the number of installed CMs of the own system 100 × Nr buffer IDs transmits the generated buffer IDs as empty reception buffer IDs to the initial notification transmission source CM. Then, the receiving master CM ends the process started in step S1006 and the initial notification.

  On the other hand, the transmission master CM that has received a plurality of empty reception buffer IDs prepares an empty reception buffer management table 160 in its own memory 112, and the received empty reception buffer management table 160 stores each received empty reception buffer ID and The total number of received empty reception buffer IDs is registered (step S1007). The empty reception buffer management table 160 prepared by the transmission master CM on the memory 112 includes only the total number of empty reception buffer IDs (the number of empty reception buffer IDs) and a plurality of empty reception buffer IDs as shown in FIG. Can be stored.

  The transmission master CM that has finished the process of step S1007 (FIG. 4A) changes the negotiation state on the group management table 150 to “negotiated” (step S1008). Furthermore, the transmission master CM sets the buffer set empty waiting state, the BS management table transfer state, and the BS periodic switching process activation state on the group management table 150 to “not waiting”, “not transferring”, “ It is initialized to “not activated” (step S1009).

  Thereafter, in step S1011 (FIG. 4B), the reception master CM sets a BS management table in which various initial values for each transmission-side buffer set n (n = 0 to Ns−1: Ns is the number of buffer sets) are set. Are prepared on the memory 112.

  Hereinafter, the BS (buffer set) management table for the transmission side buffer set and the BS management table for the transmission side buffer set n are also referred to as a transmission side BS management table and a transmission side BS management table n, respectively.

  FIG. 7 shows an example of the transmission side BS management table. As is clear from FIG. 7, the transmission-side BS management table is a table that can store a plurality of sets of transmission buffer IDs and reception buffer IDs, together with a BS number, an empty reception buffer ID waiting state, and a table transfer waiting state.

In step S1011, the reception master CM performs a process of preparing, on the memory 112, each reception-side BS management table n (n = 0 to “Nr−1”) in which the following information is set.

Buffer set number: n
Empty reception buffer ID waiting state: “Not received”
Table transfer wait state: "Not waiting"
Each transmission buffer ID: “Invalid”
Each receive buffer ID: “Invalid”

  The transmission master CM that has completed the process of step S1011 transmits an initial construction negotiation notification having a predetermined content to each slave CM in the own system 100 (step S1012).

  Thereafter, the transmission master CM initializes various information on the group management table 150 (step S1013). Specifically, the transmission master CM sets the stored BS number on the group management table 150 (FIG. 3), the BS switching state, and each switching request reception state (such as “CM10 switching request reception state”) to “0”. , “Not switching” and “Not requesting”. Further, the transmission master CM initializes the buffer set 0 use state to “in use” and initializes each other buffer set use state to “not in use”.

  In subsequent step S1014, the transmission master CM performs processing for preparing on its own memory 112 a buffer management table in which various initial values are set for each transmission-side buffer set n (n = 1 to Ns−1).

  As shown in FIG. 8, the buffer management table can store 1024 pieces of index information including a storage state, a session number, a copy destination LUN, a copy destination LBA, and the number of blocks together with a buffer set number, an index information pointer, and the like. It is a table.

  The storage state in the index information m (m = 0 to 1023) on the buffer management table is information indicating whether or not the storage of data in the m-th unit storage area of the buffer is completed. Here, the unit storage area of the buffer is a storage area (storage area of 8 KB (= 8 MB / 1024)) obtained by dividing the buffer (storage area used as a buffer) by 1024. The mth unit storage area of the buffer is the mth unit storage area counted from the head side of the buffer (the unit storage area at the head of the buffer is the 0th unit storage area). As already described, the storage area on the memory 112 that the CM 110 uses as Y buffers is “8 MB of“ default address + b · 8 MB ”(b = 0 to Y−1) as a head address. Consecutive storage areas ”. Therefore, the buffer prepared on the memory 112 of the transmission CM is also such, but the transmission CM is “8 MB of“ default address + b · 8 MB ”(b = 0 to Ns−1) as the head address. The “continuous storage area” is used as a buffer of the buffer set b.

  The copy destination LUN, the copy destination LBA, and the number of blocks in the index information m are the storage area of the storage unit 120 in the receiving system 100 to which the data stored in the mth unit storage area / the data stored thereafter is to be written. It is information which shows. The session information in the index information is identification information attached to a series of processes related to the own index information.

The buffer set number on the buffer management table is the BS number of the transmitting BS to which the own table belongs. The index information pointer is updated so that the total number of index information actually set on the buffer management table and the smallest index information number (m in index information m) in which the session number or the like is not set are indicated. Value. The buffer transfer wait status becomes "Waiting" when the status is "Waiting for setting of index information in its own buffer management table and completion of data storage in the corresponding buffer". Information to be changed.

  The reception buffer ID on the buffer management table is the buffer ID of the transfer destination buffer of the data on the buffer included in the same transmission side buffer set. The reception buffer ID is information set in the buffer management table in order to notify the reception CM of the transfer destination buffer (details will be described later).

  The process of step S1014 is a process of preparing the memory 112 with the following information set as the BS management table of each receiving-side buffer set n (n = 0 to “Nr−1”). .

Buffer set number: n
Index information pointer: 0
Receive buffer ID: “Invalid”
Buffer transfer wait state: "Not waiting"
Each storage state: “Not stored”,
Each session number: “Invalid”
Each copy destination LUN: “Invalid”
Each copy destination LBA: “Invalid”
Number of blocks: “Invalid”

  The transmission master CM that has finished the process of step S1014 prepares an empty BS switching waiting queue and a storage completion waiting queue on its own memory 112 (step S1015).

  Further, the transmission slave CM that has received the initial construction negotiation notification also initializes various information on the group management table 150, similarly to the transmission master CM (step S1021). However, the usage status of each BS is information used only by the transmission master CM. Therefore, in step S1021, the transmission slave CM sets the stored BS number, BS switching state, and each switching request state on the group management table 150 to “0”, “not switching”, and “not requesting”, respectively. Only the process of initializing to "" is performed.

  Thereafter, the transmission slave CM performs the same processes as steps S1014 and S1015 in steps S1022 and S1023, respectively.

  When the process of step S1023 is completed, the transmission slave CM notifies the transmission master CM (master CM in the own system 100) of the completion of the process.

  On the other hand, the transmission master CM that has completed the process of step S1015 waits for the process by each slave CM to be completed (step S1016).

  Then, the transmission master CM that recognizes that the processing by each slave CM has been completed notifies the main transmission master CM of the completion of the processing when the transmission master CM itself is a sub transmission master CM (step S1017). Then, the sub transmission master CM ends this initial construction negotiation process.

If the transmission master CM recognizes that the processing by each slave CM has been completed, if it is the main transmission master CM, the transmission master CM waits for notification of the completion of the processing from each sub transmission system 100 (step). S1018). And the main transmission master CM
When it is confirmed that the processing by each sub-transmission system 100 is completed, the BS regular switching processing (details will be described later) is started (step S1019), and then the initial construction negotiation processing is terminated.

  Next, the contents of the copy range setting work and the functions of the CM 110 related to the work will be described.

  First, the initial copy session management table 180 generated on the memory 112 of each transmission CM as a result of the copy range setting work will be described.

  FIG. 9 shows an example of the initial copy session management table 180. The copy source LUN, the copy source LBA, and the number of blocks on the initial copy session management table 180 are a continuous storage area (hereinafter referred to as a copy source storage area) in one logical unit on the transmission system side to be backed up. Information). The copy destination LUN, the copy destination LBA, and the number of blocks are continuous storage areas (hereinafter referred to as copy destination storage areas) in one logical unit on the receiving system side in which data in the copy source storage area is to be stored. Information to prescribe.

  The initial copy state is information indicating whether or not an initial copy process (FIG. 11; details will be described later) regarding the own table 180 is started.

The session phase is information indicating whether or not the initial copy process relating to the own table 180 has been completed. This session phase is changed to “Equivalent” when the initial copy processing relating to the own table 180 is completed. The session status is information that is changed to “Suspend” when an instruction to disconnect the reception system 100 and the transmission system 100 is given.

  During the copy range setting operation, the user uses the host 300 to make initial copy session information (copy source LUN, copy destination LUN, copy source LBA, copy destination LBA and block) for each storage area to be used as the copy source storage area. Number).

  As schematically shown in FIG. 10, the initial copy session information set by the user is transmitted from the host 300 to the master CM in the transmission system 100 including the copy source logical unit (step S1101).

  The transmission master CM that has received the initial copy session information determines whether or not the negotiation state on the group management table 150 is “negotiated” (step S1102). If the negotiation state is not “negotiated” (step S1102; NO), the transmission master CM transmits an abnormal end (information indicating that the process has ended abnormally) to the host 300 (step S1107). Then, the transmission master CM ends the process.

  When the negotiation state is “negotiated” (step S1102; YES), the transmission master CM transmits initial copy session information to the copy session handling CM (step S1103). Here, the CM in charge of the copy session is a CM 110 in charge of controlling the logical unit identified by the copy source LUN in the initial copy session information to be processed. If the transmission master CM = the CM in charge of the copy session, the processes in steps S1104 and S1005 are executed by the transmission master CM as the CM in charge of the copy session without executing the process in step S1103.

In step S1104, the CM in charge of the copy session that has received the initial copy session information generates a unique session ID (a session ID that has been generated in the past and is different from the session ID that is valid at the present time). Then, the CM in charge of the copy session sets the generated session ID, “Active”, “Copying” as the session ID, session status, session phase, initial copy session information, and initial copy state, respectively.
“, The initial copy session information received, and the initial copy session management table 180 in which“ not activated ”is set are generated on the memory 112.

  The copy session responsible CM that has finished the process of step S1104 transmits the generated initial copy session management table 180 to the copy session responsible CM in the corresponding receiving system 100 (step S1105). The CM in charge of the copy session in the corresponding receiving system 100 stores the received initial copy session management table 180 in its own memory 112 (step S1106), and then notifies the transmission master CM that the processing has been completed.

  The transmission master CM notified of the completion of the process transmits a normal end (information indicating that the process is normally completed) to the host 300 (step S1108). Thereby, the process for generating one initial copy session management table 180 on the memory 112 of one transmission CM is completed.

  Hereinafter, the contents of various processes executed by the CM 110 alone or in cooperation with other CMs 110 will be described.

<Initial copy processing>
The CM in charge of the copy session that has generated the initial copy session management table 180 on its own memory 112 starts the initial copy process of the procedure shown in FIG. Then, the copy session CM first determines whether or not the session status on the initial copy session information table is “Active” (step S1201).

  If the session status is not “Active” (step S1201; NO), the copy session responsible CM ends the initial copy process.

When the session status is “Active” (step S1201; YES), the CM in charge of the copy session determines whether the session phase is “Copying” (step S1202). If the session phase is not “Copying” (step S1202; NO), the copy session CM ends the initial copy process.

If the session phase is “Copying” (step S1202; YES)
The CM in charge of the copy session determines whether or not the initial copy state is “not activated” (step S1203). When the initial copy state is “not activated” (step S1203; YES), the copy session managing CM prepares a copy source bitmap for copy target management on its own memory 112 (step S1204). Here, the copy source bitmap for copy target management is data (bitmap data) in which each bit indicates whether or not data on a specific storage area is to be copied.

The CM in charge of the copy session that has completed the process of step S1204 changes the initial copy state on the initial copy session management table 180 to “activated” (step S1205). Next, the CM in charge of the copy session copies “a part of the data to be copied in the copy source bitmap to the copy destination, and inverts the value of the bit in the copy source bitmap relating to the copied data. "Partial copy processing" is performed (step S1206).

  The CM in charge of the copy session that has completed the partial copy process determines whether or not the initial copy (copy of the content specified in the initial copy session information table 180) has been completed (step S1207). If the initial copy has not been completed (step S1207; NO), the copy session CM sets the execution time (date and time) of the next initial copy process (step S1208) and then sets the initial initial The copy process ends.

  As is apparent from the above description, the initial copy state is “started up” when the initial copy process is executed for the second and subsequent times. Therefore, the CM in charge of the copy session branches to the “NO” side in step S1203. In other words, the CM in charge of the copy session executes the partial copy process in step S1206 without performing the processes in steps S1204 and S1205.

  When the initial copy is completed as a result of repeating the above processing (step S1207; YES), the copy session responsible CM changes the session phase on the initial copy session management table 180 to “Equivalent” (step S1209). ). Then, the copy session responsible CM transmits an initial copy completion notification to the copy session responsible CM in the corresponding receiving system (step S1210).

  The CM in charge of the copy session in the corresponding receiving system that has received this notification changes the session phase to “Equivalent” (step S1211), and then notifies the CM in charge of the copy session in the transmission system of the completion of the process.

《Write request analysis / response processing》
The write request analysis / response process is a process of the procedure shown in FIG. 12 executed by the transmission CM that has received the write request from the host 300. As already described, the host 300 transmits a write request for a certain logical unit to the transmission CM in charge of that logical unit. Therefore, it is the CM in charge of the copy session that executes the write request analysis / response process.

  As shown in FIG. 12, the CM in charge of the copy session that has received the write request, first, the write request is a write request to one of the copy ranges (copy source storage area) that it is in charge of. Whether or not (step S1301). More specifically, in step S1301, the CM in charge of the copy session has an initial copy session management table 180 (FIG. 9) in which the write range of the received write request data overlaps with the copy source storage area in the memory 112. Judge whether or not. If the initial copy session management table 180 exists, the copy session responsible CM stores the copy destination LUA set on the initial copy session management table 180 as the transfer destination LUA. Further, the CM in charge of the copy session obtains the LBA in the copy destination LU corresponding to the LBA for which the write request is made based on the copy source LBA and the copy destination LBA set in the initial copy session management table 180. Is stored as the transfer destination LBA.

If the received write request is not a write request to any of the copy ranges (step S1301; NO), the copy session responsible CM performs only a process in response to the received write request (step S1302). Here, the process of responding to the received write request is a process of writing the data of the content requested by the write request to the storage unit 120 and then transmitting an end notification to the host 300. is there.

  Then, the CM in charge of the copy session responding to the write request ends the write request analysis / response process.

  On the other hand, if the received write request is a write request to one of the copy ranges (step S1301; YES), the CM in charge of the copy session indicates that the status and phase of the initial copy session related to the copy range are “ It is determined whether or not “Active” or “Equivalent” is set (step S1303).

As already described, the phase of an initial copy session does not become “Equivalent” unless the copy in the range specified by the user is completed. Therefore, if the phase of the range overlap initial copy session where the write range of the write request overlaps with the copy range is not “Equivalent”, the data copy has not been completed to the level where it is necessary to guarantee the order of data write. It will be. In addition, the status of a certain initial copy session becomes “Suspend” when the user gives an instruction to disconnect the receiving system. Therefore, even when the status of the range duplication initial copy session is not “Active”, it is not necessary to guarantee the order of data writing.

  Therefore, if status = “Active” and phase = “Equivalent” have not been established (step S1303; NO), the copy session CM performs copy source bitmap update processing (step S1304). Here, the copy source bitmap update process is a process of updating the copy source bitmap so that the write range of the write request being processed is copied in the initial copy process. This copy source bitmap update process is named “update process”, but if the write range of the write request being processed is a range that has not yet been copied in the initial copy process, the copy source bit The process does not change the contents of the map.

  The CM in charge of the copy session that has finished the process of step S1304 responds to the write request (step S1305). Then, the copy session CM ends this write request analysis / response process.

  On the other hand, if the status = “Active” and the phase = “Equivalent” are established (step S1303; YES), the copy session CM activates the storage reservation / queuing process for the received write request (step S1306). . Note that “a certain CM 110 activates a process” means “the CM 110 starts a process as a separate process without interrupting the process being executed”.

  Then, the copy session CM ends this write request analysis / response process.

<< Storage reservation / queuing process >>
The storage reservation / queuing process is a process of the procedure shown in FIG. In the following description, Nc is the buffer set number on the group management table 150 at that time. The buffer Nc and the buffer management table Nc are a buffer and a buffer management table for the transmission-side buffer set Nc.

  As shown in FIG. 13, the CM in charge of the copy session that has started the storage reservation / queuing process is first switching the buffer set (the BS switching state on the group management table 150 is “switching”). ) Is determined (step S1401).

  If the buffer set is being switched (step S1401: YES), the copy session CM places the write request in the BS switching queue (step S1404). Then, the copy session CM ends this storage reservation / queuing process.

  On the other hand, if the buffer set is not being switched (step S1401; NO), the copy session CM determines whether or not the empty size of the buffer Nc is equal to or larger than the size of the current write data (step S1403). More specifically, the CM in charge of the copy session calculates the free size “(1024-m) × 8 KB” of the buffer Nc from the value m of the index information pointer on the buffer management table Nc. Then, the CM in charge of the copy session determines whether or not the calculated empty size of the buffer Nc is equal to or larger than the size of the current write data.

  The contents of the processing executed when the free size of the buffer Nc is less than the size of the write data is whether the CM in charge of the copy session executing this storage reservation / queuing processing is the main transmission master CM. It depends on. The CM in charge of the copy session, which is the slave CM or the sub transmission master CM, starts buffer set switching request processing (details will be described later) when the empty size of the buffer Nc is less than the size of the write data (step S1402; NO). (Step S1404). Further, the copy session handling CM, which is the master CM, activates a buffer set switching process (details will be described later) when the empty size of the buffer Nc is less than the size of the write data (step S1402; NO) (step S1404). .

  The CM in charge of the copy session that has completed the processing in step S1404 places the write request in the buffer set switching queue (step S1404), and then ends this storage reservation / queuing processing.

  On the other hand, when the empty size of the buffer Nc is equal to or larger than the size of the write data (step S1402; YES), the CM in charge of the copy session stores unstored same-range index information on any buffer management table related to the buffer set in use. It is determined whether or not it exists (step S1405). Here, the buffer management table related to the in-use buffer set is a buffer set whose use state (“buffer set n use state”) on the group management table 150 is “in use”. In addition, the unstored same range index information (see FIG. 8) indicates that the storage state is “unstored” and the storage area indicated by the copy destination LUN, the copy destination LBA, and the number of blocks is the write data of this time. This is index information that overlaps the range.

  If unstored same-range index information does not exist (step S1405; NO), the copy session CM performs index information update processing (step S1407).

  During this index information update process, the copy session CM first generates a unique session number Snew. Next, the CM in charge of the copy session stores the size of the write data as the current data size, and stores the current index information pointer values on the buffer management table Nc as m and m0. Then, the following processing is performed using the transfer destination LUN and transfer destination LBA whose values are determined during the processing (determination) in step S1301.

  The copy session CM first sets Snew, transfer destination LUN, transfer destination LBA, & H10 as the session number, LUN, LBA, and number of blocks of the index information m, and sets the transfer destination LUN, etc., set in the buffer management table Nc. Process.

Next, the CM in charge of the copy session performs a process of incrementing the value of the index information pointer on the buffer management table Nc by “1” and the following calculation process.
m = m + 1
Current data size = Current data size-8KB
Transfer destination LBA = Transfer destination LBA + & H10

  Thereafter, the copy session CM determines whether or not the current data size> 8 KB is satisfied. Then, if the current data size> 8 KB is established, the copy session CM executes again the processing after the above-described transfer destination LUN etc. setting processing.

  On the other hand, when the current data size ≦ 8 KB is established, the CM in charge of the copy session sets the current data size in the unit of Snew, transfer destination LUN, transfer destination LBA, and block number, respectively, the session number, LUN, LBA, and number of blocks. Is set in the buffer management table Nc.

  Then, the copy session responsible CM increments the index information pointer value and the m value on the buffer management table Nc by “1”, respectively, and then ends the index information update process (step S1407).

  The copy session managing CM that has finished the index information update process stores the write request copy control table for allowing the write data to be read later from the storage unit 120 and stored in the storage areas after the unit storage area m0 of the buffer Nc. It is generated on 112 (step S1408).

  More specifically, the CM in charge of the copy session has the configuration shown in FIG. 14 and generates a write request copy control table 190 in which the following information is set as various information on the memory 112.

Write request copy control name: Automatically generated unique name Copy source LUN: LUN of write data write destination logical unit
Copy source LBA: address (LBA) of write data write start logical block
Number of blocks: Number of blocks of write data Buffer set number: Nc
Storage area offset: m0
Number of index information: m-m0
Session number: New

  Then, the CM in charge of the copy session responds to the write request (step S1409), starts the data storage process for the generated write request copy control table 190 (step S1410), and then ends the storage reservation / queuing process. .

<Data storage processing>
The data storage process is a process of the procedure shown in FIG.

That is, the copy session CM that has started the data storage process for a certain write request copy control table 190 first reads the storage reservation data from the storage unit 120 and stores it in the buffer based on the information on the table 190 ( Step 1501) is performed. More specifically, the CM in charge of the copy session reads data stored in the storage area indicated by the copy source LUN, the copy source LBA, and the number of blocks on the write request copy control table 190 from the storage unit 120. If the BS number and the storage area offset on the write request copy control table 190 are expressed as n1 and n2, respectively, the CM in charge of the copy session stores the read data in the storage areas after the unit storage area n2 of the buffer n1. Remember.

  The CM in charge of the copy session that has completed the processing in step 1501 changes the storage state in one or a plurality of index information related to the data stored in the buffer to “in storage” (step S1502).

  Thereafter, the copy session CM determines whether or not the storage completion waiting queue is empty (step S1503). When the storage completion waiting queue is not empty (step S1503; YES), the copy session responsible CM determines whether or not the write request for which storage reservation is possible is in the storage completion waiting queue (step S1504). If the write request for which storage reservation is possible is in the storage completion waiting queue (step S1504; YES), the write request is taken out of the storage completion waiting queue, and the storage reservation / queue for the extracted write request is stored. Ing process (FIG. 13) is activated (step S1505).

  The copy session responsible CM that has finished the processing of step S1505 determines whether or not the buffer transfer waiting state on the buffer management table is “waiting” (step S1506).

  If the storage completion waiting queue is empty (step S1503; YES), the copy session responsible CM performs the process (judgment) in step S1506 without performing the process in step S1505. In addition, the CM in charge of the copy session does not perform the process of step S1505 and does not perform the process of step S1506 even when the write request for which storage reservation is possible is not queued in the storage completion waiting queue (step S1504; NO). Process.

  If the buffer transfer waiting state on the buffer management table is “waiting” (step S1506; YES), the CM in charge of the copy session determines whether the entire storage state on the buffer management table is “stored”. Is determined (step S1506; YES). Then, the CM in charge of the copy session ends this storage processing when some storage states on the buffer management table are “not stored” (step S1507; NO). The CM in charge of the copy session also ends this storage process when the transfer waiting state on the buffer management table is not “waiting” (step S1506; NO).

  If the storage status on the buffer management table is “stored” (step S1507; YES), the copy session CM changes the transfer waiting status on the buffer management table to “not waiting”. (Step S1508). Then, the copy session CM activates the buffer management table and the buffer transfer process (FIGS. 21A to 21G) (step S1508).

《Buffer set switching request processing》
FIG. 16 shows a flowchart of BS switching request processing.

  As already described, this BS switching request process is a process executed by the copy session responsible CM that is the slave CM or the sub transmission master CM when the empty size of the buffer Nc is less than the size of the write data. .

As shown in FIG. 16, the CM in charge of the copy session that has started the BS switching request process first determines whether or not the BS switching request status on the group management table 150 is “requesting” (step). S1601).

  If the BS switching request status is not “requesting” (step S1601; NO), the copy session CM changes the BS switching request status on the group management table 150 to “requesting” (step S1602). Next, the copy session responsible CM transmits a buffer set switching request in which the new buffer set number Nnew is set to the main transmission master CM (step S1603). Here, the new buffer set number Nnew is a number of 0 or more and Ns−1 or less that is different from Nc and is generated by a predetermined algorithm based on the buffer set number Ns and the storage buffer set number Nc at that time. (Integer value). As the “predetermined algorithm”, for example, when “Nc is“ Ns−1 ”, the new buffer set number Nnew is set to“ 0 ”, and Nc is not“ Ns−1 ”. In this case, the new buffer set number Nnew is “Nc + 1” ”.

  Then, the CM in charge of the copy session that has finished transmitting the buffer set switching request (step S1603) ends this buffer set switching request processing.

  On the other hand, when the BS switching request status is “in request” (step S1601; NO), the CM in charge of the copy session ends this buffer set switching request processing without performing any particular processing.

《Buffer set switching processing》
The buffer set switching process is the procedure shown in FIGS. 17A and 17B that is started by the main transmission master CM when the buffer set switching request is received and when the empty size of the buffer Nc is less than the size of the write data. It is processing of. When the empty size of the buffer Nc is less than the size of the write data, the main transmission master CM generates (determines) a new buffer set number Nnew number and then starts this buffer set switching process.

  As shown in FIG. 17A, the main transmission master CM that has started the buffer set switching process first determines whether the new BS number and the stored BS number on the group management table 150 match. (Step S1701).

  If both numbers match (step S1701; NO), the buffer set has already been switched to the new buffer set. Therefore, the main transmission master CM transmits a switching completion notification to the switching request source CM (step S1702), and ends this buffer set switching processing.

  As shown in FIG. 18, the CM in charge of the copy session that has received the switching completion notification extracts the write request from the BS switching waiting queue and starts storage reservation / queuing processing for the extracted write request (step S1801). . Then, the CM in charge of the copy session changes the BS switching request status on the group management table 150 to “not being requested” (step S1802), and then ends the processing in this figure.

  Returning to FIG. 17A, the description of the buffer set switching process will be continued.

  If the new BS number and the stored BS number do not match (step S1701; NO), the main transmission master CM changes the switching request reception state related to the switching request source CM on the group management table 150 to “receiving”. (Step S1703).

  Next, the main transmission master CM determines whether or not a switching request from another CM (a CM 110 different from the current switching request source CM) has been received based on the switching request reception state on the group management table 150. (Step S1704).

  If a switching request from another CM has been received (step S1704; YES), the processing in FIG. 17B (the main part of the buffer set switching processing) has already started. Therefore, the main transmission master CM ends this BS switching process.

  If a switching request from another CM has not been received (step S1704; NO), the main transmission master CM determines whether there is an unused buffer set (step S1705).

  Here, the unused buffer set is a buffer set that does not hold any data to be transferred to the receiving system 100 (that is, a buffer set that can be used as a new buffer set). Although details will be described later, the transmission master CM updates the group management table 150 so that the use state (“buffer set n use state”) regarding the unused buffer set becomes “unused” (see FIG. 23B). . Therefore, the main transmission master CM determines whether or not there is an unused buffer set based on the presence or absence of a buffer set usage state that is “unused”. If there is no unused buffer set (step S1705; NO), the main transmission master CM temporarily ends this buffer set switching process because there is no buffer set that can be used as a new buffer set. Then, when an unused buffer set is prepared by the release process (see FIG. 23B), the main transmission master CM starts the buffer set switching process again.

  When there is an unused buffer set (step S1705; YES), the main transmission master CM instructs the other transmission CMs (main transmission slave CM, sub transmission CM) to change the BS switching state with a predetermined content. Is transmitted (FIG. 17B; step S1710).

  The main transmission master CM that has finished transmitting the BS switching state change instruction changes the buffer set switching state on the group management table 150 to “switching” (step S1711).

  Each transmission CM that has received the BS switching state change instruction also changes the buffer set switching state on the group management table 150 to “switching” (step S1711). Thereafter, each transmission CM notifies the main transmission master CM of the completion of processing.

  The main transmission master CM that has finished changing the BS switching state waits for the processing of each of the other transmission CMs (change of the BS switching state to “switching”) to be completed (step S1712).

  When the main transmission master CM finds that the processing by each of the other transmission CMs has been completed, the main transmission master CM transmits a BS switching request in which the new BS number is set as a parameter to each of the other transmission CMs (step S1713).

  Next, the main transmission master CM changes the stored BS number and BS switching state on the group management table 150 to the new BS number and “not switching” (step S1714). Further, the main transmission master CM changes the BS use state (corresponding BS use state in FIG. 17B) related to the new buffer set on the group management table 150 to “in use” (step S1714).

Each transmission CM that has received the buffer switching request also performs the same processing (step S1714) as that performed by the main transmission master CM.

  The sub-transmission master CM that has finished the process of step S1714 asynchronously starts the empty reception buffer ID assignment process for the old buffer set (the buffer set identified by the BS number before the change) (step S1716). Here, “a certain CM 110 activates a process asynchronously” means “when the load of the CM 110 is reduced regardless of the process being executed at all times, etc. Is to start.

  Then, the sub transmission master CM transmits a process completion notification to the main transmission master CM, and then ends the process.

  In addition, each transmission slave CM that has completed the process of step S1714 transmits a process completion notification to the main transmission master CM without performing the process of step S1716, and then ends the process.

  On the other hand, the main transmission master CM that has finished the process of step S1714 starts the empty reception buffer ID assignment process for the old buffer set (step S1715). Thereafter, the main transmission master CM waits for the processing of each other transmission CM to be completed (step S1717). Then, when the main transmission master CM can confirm that the processing by the other transmission CMs has been completed, the main transmission master CM transmits to each switching request source CM whose switching request state on the group management table 150 is “requesting”. On the other hand, after the switching completion notification is transmitted (step S1718), the BS switching process is terminated.

<< Available reception buffer ID assignment process >>
FIG. 19 shows a flowchart of the empty reception buffer ID assignment process.

  As already described, this empty reception buffer ID assignment process is performed for the old buffer set (the buffer set for which index information has been written to the buffer management table).

  As shown in FIG. 19, the transmission master CM that has started the empty reception buffer ID assignment processing first determines whether or not the necessary number of empty reception buffer IDs are stored in the empty reception buffer management table 160 ( Step S1901). Here, the necessary number of empty reception buffer IDs is the same number of empty reception buffer IDs as the number of CMs L installed in the own system.

  If the necessary number of empty reception buffer IDs are not stored in the empty reception buffer management table 160 (step S1901), the transmission master CM changes the waiting state of the empty reception buffer ID on the group management table 150 to “waiting”. (Step S1907). Then, the transmission master CM ends this empty reception buffer ID assignment process.

  On the other hand, when the necessary number of empty reception buffer IDs are stored in the empty reception buffer management table 160 (step S1901; YES), the transmission master CM sets the empty reception buffer IDs stored in the empty reception buffer management table 160. An empty reception buffer ID to be assigned to each transmission buffer ID related to the old buffer set is selected from among them (step S1902).

Next, the transmission master CM sets each transmission buffer ID in the BS management table for the old buffer set, and uses the empty reception buffer ID selected for each transmission buffer ID as the reception buffer ID corresponding to each transmission buffer ID. Set (step S190
3).

  Thereafter, the transmission master CM deletes the selected empty reception buffer ID from the empty reception buffer management table 160 and changes the number of empty reception buffer IDs on the empty reception buffer management table 160 to a value representing the current state ( Step S1904) is performed.

  Next, the transmission master CM starts the BS management table transfer process for the old buffer set asynchronously (step S1905). Then, the transmission master CM starts the buffer management table and buffer transfer processing for the old buffer set asynchronously (step S1906), and then ends this empty reception buffer ID allocation processing.

<< Buffer set management table transfer process >>
20A to 20D are flowcharts of BS management table transfer processing. In the drawings used for the description below and in the following description, X is a BS number of a buffer set (a BS number of an old buffer set or the like) that is a processing target of a certain process.

  As shown in FIG. 20A, the transmission master CM that has started the BS management table transfer process first determines whether or not the BS management table transfer state on the group management table 150 is “in transfer” (step S21). S2101).

  When the BS management table transfer state is “transferring” (step S2101; YES), the transmission master CM is “waiting” for the transfer waiting state on the BS management table X (the BS management table belonging to the buffer set X). (Step S2108). Then, the transmission master CM ends this BS management table transfer process.

  On the other hand, if the BS management table transfer state is not “transferring” (step S2101; NO), the transmission master CM changes the transfer waiting state on the BS management table X to “transferring” (step S2102). Then, the transmission master CM transmits all information on the BS management table X (hereinafter referred to as management table information) to the master CM in the corresponding reception system (step S2103).

  Hereinafter, the operation of the reception master CM that has received the management table information will be described separately for the case where the corresponding reception system 100 is the sub reception system 100 and the case where the corresponding reception system 100 is the main reception system 100.

  When the corresponding receiving system 100 is the sub-receiving system 100, the master CM in the corresponding receiving system 100 (that is, the sub-receiving master CM) executes the process of the procedure shown in FIG. 20B.

  That is, the sub-reception master CM first adds each transmission buffer ID · in the management table information to the BS management table specified by the BS number in the received management table information (hereinafter referred to as a corresponding BS management table). A reception buffer ID pair is set (step S2111).

  Next, the sub reception master CM changes the buffer ID reception state (“not received” at this time) on the corresponding BS management table to “received” (step S2112).

  Thereafter, the sub reception master CM determines whether or not the expanded BS number on the group management table 150 is “invalid” (a value set as an initial value by the initial construction negotiation process) (step S2113).

  If the expanded BS number is “invalid” (step S2113; YES), the sub reception master CM changes the expanded BS number on the group management table 150 to the BS number in the received management table information (step S2114). ). Then, the sub reception master CM transmits a BS management table reception completion notification to the transmission source CM of the management table information (step S2115).

  Further, if the expanded BS number is not “invalid” (step S2113; NO), the sub reception master CM sends a BS management table reception completion notification to the management table information transmission source without performing the process of step S2114. The message is transmitted to the CM (step S2115).

  Thereafter, the sub reception master CM determines whether or not the reception of all the element information related to the buffer set X is completed (step S2121). Here, the element information related to the buffer set X is held in each component (BS management table, several sets of buffer management tables and buffers) of the transmission-side buffer set identified by the BS number in the received management table information. It is the information that has been done.

  If reception of all element information related to the buffer set X has been completed (step S2121; YES), the sub reception master CM transmits a BS reception completion notification to the main transmission master CM (step S2122), and then ends the process. To do. Further, if the reception of all the element information related to the buffer set X has not been completed (step S2121; NO), the sub reception master CM ends the process without transmitting the BS information reception completion notification.

  When the corresponding receiving system 100 is the main receiving system 100, the master CM (that is, the main receiving master CM) in the corresponding receiving system 100 that has received the management data information executes the processing of the procedure shown in FIG. 20C.

  That is, the main reception master CM that has received the management data information performs the same processing as the processing in steps S21111 to S2115 described above in steps S2131 to S2135.

  Then, the main reception master CM that has finished the process of step S2135 determines whether or not the reception of all the element information regarding the buffer set X has been completed (step S2141). If reception of all element information related to the buffer set X has not been completed (step S2141; NO), the main reception master CM ends the process of FIG. 20C without performing any particular process.

  On the other hand, when reception of all element information related to the buffer set X has been completed (step S2141; YES), the main reception master CM sets the reception status of the reception system 0 related to the buffer set X on the group management table 150 to “Received”. "(Step S2142).

  Thereafter, the main reception master CM determines whether or not the reception states of the respective systems 100 related to the buffer set X are all “received” (step S2143). The main reception master CM starts the expansion processing for the buffer set X asynchronously when all the reception states of the respective systems 100 regarding the buffer set X are “received” (step S2143; YES). After step S2144), the process ends.

If the reception status of each system 100 related to the buffer set X is not all “received” (step S2143; NO), the main reception master CM ends the process without performing the process of step S2144. To do.

  Returning to FIG. 20A, the operation of the transmission master CM after receiving the BS management table reception completion notification will be described.

  The transmission master CM that has received the BS management table reception completion notification changes the BS management table transfer state on the group management table 150 to “not transferring” (step S2104).

  Next, the transmission master CM determines whether or not there is a BS management table (in FIG. 20A, “transfer waiting table”) whose table transfer wait state is “waiting” (step S2105). If there is a BS management table whose table transfer wait state is “waiting” (step S2105; YES), the transmission master CM starts the BS management table transfer process for the BS management table. After (step S2106), the BS management table transfer process that has been executed is terminated.

  Here, the operation of the main transmission master CM that has received the BS reception completion notification will be described with reference to FIG. 20D.

  When the BS reception completion notification is received, the main transmission master CM executes the process of the procedure shown in FIG. 20D. That is, the main reception master CM that has received the BS reception completion notification first sets the reception state (buffer set X transmission system 0 reception state) regarding the buffer set X and the own system 100 on the group management table 150 to “Received”. Change (step S2151). Next, the main transmission master CM determines whether or not all the reception statuses for each system related to the buffer set X are “received” (step S2152). If all the reception statuses of the buffer set X for each system are “Received” (step S2152; YES), the main transmission master CM activates the expansion process for the buffer set X (step S2153). Then, the processing in this figure is terminated. Further, when only a part of the reception status of each system regarding the buffer set X is “Received” (step S2152; NO), the main reception master CM performs the processing shown in FIG. Exit.

<< Buffer management table and buffer transfer process >>
21A to 21G show a flowchart of the buffer management table and buffer transfer processing.

  The buffer management table and the buffer transfer process are processes started by the transmission master CM. The transmission master CM that has started the buffer management table and buffer transfer processing for the buffer set X notifies each CM in the own system 100 to transfer the buffer management table by notifying the BS number X and the assigned reception buffer ID ( Step S2211).

  Here, the allocation reception buffer ID is the reception buffer ID stored in the BS management table X (BS management table of the buffer set X) in association with the buffer ID (transmission buffer ID) of the transfer target buffer. That is.

Each transmission slave CM notified of the allocation reception buffer ID and the like sets the allocation reception buffer ID as the reception buffer ID in the buffer management table X (step S2212). Also, the transmission master CM that has finished the process of step S2211 sets the assigned reception buffer ID as the reception buffer ID in the buffer management table X (step S2212).
.

  Next, each transmission CM determines whether or not all storage states in the index information 0 to m-1 (m is the value of the index information pointer at that time) on the buffer management table X are “stored”. Judgment is made (step S2213). That is, each transmission CM determines whether or not all the storage reservation data is stored in the transfer target buffer (step S2213).

  If there is storage reservation data that has not been stored in the transfer target buffer (step S2213; NO), each transmission CM changes the buffer transfer waiting state on the buffer management table X to “waiting”. (Step S2213). Then, each transmission CM ends this buffer management table and buffer transfer processing.

  On the other hand, when the storage of all the storage reservation data has been completed in the transfer target buffer (step S2213; YES), each transmission CM uses the buffer management table X (all information on the buffer management table X) as the corresponding reception CM. Transmit (step S2215).

  The receiving CM that has received the buffer management table X stores the received table in the memory 112 as it is (step S2216). Then, the reception CM notifies the corresponding transmission CM of the completion of reception of the buffer management table (step S2217).

  Hereinafter, the operation of the reception CM after the processing in step S2217 will be described separately for the case where the reception CM is a slave CM and the case where the reception CM is a master CM.

  When the reception CM is a slave CM, the reception CM (reception slave CM) that has finished the process of step S2217 executes the process of the procedure shown in FIG. 21B.

  That is, the reception slave CM notifies the master CM in the own system of the completion of reception of the buffer management table (step S2231), and then ends the process started by reception of the buffer management table.

  On the other hand, the reception master CM notified of the reception completion of the buffer management table changes the buffer management table reception state related to the reception completion notification source CM on the BS management table X to “received” (step S2231). Then, the reception master CM ends the process.

  If the received CM is a master CM, the received CM (received master CM) that has finished the process of step S2217 executes the process of the procedure shown in FIG. 21C.

  That is, the reception master CM changes the buffer management table reception state regarding the own CM 110 on the BS management table X to “received” (step S2235). Then, the reception master CM ends the process.

  Returning to FIG. 21A, the description of the buffer management table and buffer transfer processing will be continued.

  The transmission slave CM notified of the completion of reception of the buffer management table notifies the transmission master CM of completion of transfer of the buffer management table (step S2218).

  The transmission master CM notified of the completion of the buffer management table transfer requests each CM 110 in its own system 100 to transfer the buffer by notifying the BS number X and the allocation reception buffer ID (step S2221).

  The transmission CM requested to transfer the buffer (transmission master CM, transmission slave CM) transmits the allocated reception buffer ID and the data on the buffer X (hereinafter referred to as buffer data) to the corresponding reception CM (step S2222). ).

  The receiving CM that has received the allocation reception buffer ID and the buffer data stores the received buffer data in the reception buffer identified by the received allocation reception buffer ID (step S2223). Then, the reception CM notifies the corresponding transmission CM of the completion of reception (step S2224).

  Hereinafter, the operation of the reception CM after the processing in step S2224 will be described for each type of reception CM.

  When the reception CM is a sub reception slave CM, the reception CM (sub reception slave CM) that has finished the process of step S2224 executes the process of the procedure illustrated in FIG. 21D.

  In other words, the sub reception slave CM notifies the master CM in the own system of the completion of reception of the buffer management table (step S2241), and then ends the process.

  On the other hand, the sub reception master CM notified of the reception completion of the buffer management table changes the buffer reception state regarding the reception completion notification source CM on the BS management table X to “received” (step S2242). Thereafter, the sub reception master CM determines whether or not the reception of all the element information related to the buffer set X has been completed (step S2243).

  If reception of all element information related to the buffer set X has been completed (step S2243; YES), the sub reception master CM transmits a BS reception completion notification to the main transmission master CM (step S2243), and then ends the process. To do. Further, if the reception of all the element information related to the buffer set X has not been completed (step S2243; NO), the sub reception master CM ends the process without transmitting the BS information reception completion notification.

  When the reception CM is the main reception slave CM, the reception CM (main reception slave CM) that has finished the process of step S2224 executes the process of the procedure shown in FIG. 21E.

  That is, the main reception slave CM notifies the master CM in the own system 100 of the completion of reception of the buffer management table (step S2251), and then ends the process.

  The main reception master CM notified of the completion of reception of the buffer management table changes the buffer reception state regarding the reception completion notification source CM on the BS management table X to “received” (step S2252). Thereafter, the main reception master CM determines whether or not the reception of all the element information related to the buffer set X has been completed (step S2253).

  When reception of all the element information related to the buffer set X has been completed (step S2253; YES), the main reception master CM displays the reception status regarding the buffer set X and its own system (reception system 0) on the group management table 150. It is changed to “Received” (step S2254). After that, the main reception master CM determines whether or not all the reception statuses for each system regarding the buffer set X are “received” (step S2255).

Then, the main reception master CM ends the process when the reception status of each system regarding the buffer set X is not “received” (step S2255; NO).

  Further, when the reception status for each system regarding the buffer set X is “received” (step S2255; YES), the main reception master CM activates the expansion processing for the buffer set X (step S2256). Then, the process ends.

  When the reception CM is a sub reception master CM, the reception CM (sub reception master CM) that has finished the process of step S2224 executes the process of the procedure shown in FIG. 21F.

  That is, the sub reception master CM changes the buffer reception state regarding the reception completion notification source CM on the BS management table X to “received” (step S2262). Thereafter, the sub reception master CM determines whether or not reception of all the element information regarding the buffer set X is completed (step S2263).

  If reception of all element information related to the buffer set X has been completed (step S2263; YES), the sub reception master CM transmits a BS reception completion notification to the main transmission master CM (step S2264), and then ends the process. To do. In addition, if the reception of all the element information related to the buffer set X has not been completed (step S2263; NO), the sub reception master CM ends the process without transmitting the BS information reception completion notification.

  When the reception CM is the main reception slave CM, the reception CM (main reception slave CM) that has finished the process of step S2224 executes the process of the procedure shown in FIG. 21G.

  In other words, the main reception slave CM changes the buffer reception state related to the reception completion notification source CM on the BS management table X to “received” (step S2272). Thereafter, the main reception master CM determines whether or not reception of all element information relating to the buffer set X has been completed (step S2273).

  When reception of all the element information related to the buffer set X has been completed (step S2273; YES), the main reception master CM indicates the reception status regarding the buffer set X and its own system (reception system 0) on the group management table 150. It is changed to “Received” (step S2274). Thereafter, the main reception master CM determines whether or not all the reception statuses of the buffer sets X for each system are “received” (step S2275).

  Then, the main reception master CM ends the process when the reception status of each system regarding the buffer set X is not “received” (step S2275; NO). Further, when the reception status of each system regarding the buffer set X is “received” (step S2275; YES), the main reception master CM activates the expansion processing for the buffer set X (step S2276). Then, the process ends.

《Deployment processing》
22A to 22C are flowcharts of the expansion process.

  The main reception master CM that has started expansion processing for a certain buffer set X first determines whether or not the expansion BS number X required matches the expansion BS number on the group management table 150 (step S2301).

Then, if the required BS number coincides with the expanded BS number (step S2301), the main reception master CM specifies the buffer X in the form of designating the BS number X to the master CM of each other receiving system 100. Is requested to be developed (step S2303).

  The sub-reception master CM requested to expand the buffer X requests each slave CM in the own system 100 to expand the buffer X in the form of specifying the BS number X as shown in FIG. S2311). Receiving this request, each sub reception slave CM stores each data on the buffer X in a storage area in the storage unit 120 indicated by each index information on the buffer management table X (step S2321). Thereafter, the sub reception slave CM transmits a deployment completion notification to the master CM in the own system 100 (step S2322).

  In addition, the sub reception master CM that requests each sub reception slave CM to expand the buffer X also stores each data on the buffer X in a storage area in the storage unit 120 indicated by each index information on the buffer management table X. Store (step S2312).

  The sub-reception master CM that has finished the process of step S2312 monitors the completion of the expansion of the buffer X by all CMs in the own system 100 (step S2313; NO). Then, when the buffer expansion by all CMs is completed (step S2313; YES), the sub reception master CM transmits a expansion completion notification to the main transmission master CM (step S2314), and then ends the process.

  Further, the main reception master CM that has finished the process of step S2303 (FIG. 22A) starts the process of the procedure shown in FIG. 22C.

  That is, the main reception master CM requests each slave CM in the own system 100 to expand the buffer X in the form of specifying the BS number X (step S2331). Receiving this request, each sub reception slave CM stores each data on the buffer X in a storage area in the storage unit 120 indicated by each index information on the buffer management table X (step S2341). Thereafter, the sub reception slave CM transmits a deployment completion notification to the master CM in the own system (step S2342).

  The main reception master CM that has requested each sub reception slave CM to expand the buffer X also stores each data on the buffer X in a storage area in the storage unit 120 indicated by each index information on the buffer management table X. Store (step S2332).

  The main receiving master CM that has finished the processing of step S2332 waits for completion of the expansion of the buffer X by all the CMs in the own system 100 by monitoring the transmission completion notification transmitted (step S2333). NO). Then, when the development of the buffer by all CMs is completed (step S2333; YES), the main reception master CM waits for the completion of the development of the buffer X by all the reception systems 100 (FIG. 22A: step S2305; NO).

  When the expansion of the buffer X by all the receiving systems 100 is completed (step S2305; YES), the main transmission master CM changes the expansion buffer set number on the group management table 150 in the same procedure as the generation procedure of the new buffer set number. (Step S2306).

  Thereafter, the main reception master CM determines whether or not the waiting state on the buffer management table of the next buffer set (the buffer set having the changed development buffer set number) is “waiting” (step S2307). ).

  If the development waiting state on the buffer management table of the next buffer set is “waiting” (step S2307; YES), the main reception master CM activates the development processing for the buffers in the buffer set (step S2308). ).

  Then, the main reception master CM activates a release process for the buffer set that has been developed (step S2309), and ends the process.

《Release processing》
FIG. 23A and FIG. 23B are flowcharts of the expansion process.

  As shown in FIG. 23A, the main reception master CM that has started the release process for a certain buffer set X first transmits a BS release request in which the BS number X is set as a parameter to the main transmission master CM (step S31). S2501)

  The main reception master CM transmits a BS release request to the master CM in each other reception system (step S2502).

  After that, the main reception master CM changes all the reception statuses by system regarding the buffer set X on the group management table 150 to “not received” (step S2503). Next, the main reception master CM initializes the BS management table X (step S2504). Specifically, in step S2504, the main reception master CM initializes the buffer set reception state and the expansion wait state on the BS management table X to “not received” and “not waiting”, respectively. . In step S2504, the main reception master CM sets each reception buffer ID, each transmission buffer ID, each buffer management table reception state, and each buffer reception state on the BS management table X to “invalid” and “invalid”, respectively. ”,“ Not received ”, and“ not received ”.

  Each sub-reception master CM that has received the BS release request also changes all reception statuses for each buffer set X in the group management table 150 to “unreceived” (step S2513), and initializes the BS management table X. (Step S2514).

  On the other hand, as shown in FIG. 23B, the main transmission master CM that has received the BS release request first sets each reception buffer ID set on the BS management table X in the empty reception buffer table in step S2601. sign up. The main transmission master CM also performs processing for updating the number of empty buffer IDs on the empty reception buffer table.

  Next, the main transmission master CM initializes the BS management table X (step S2602), and then transmits a BS release request in which the BS number X is set as a parameter to the sub transmission master CM (step S2603). The sub transmission master CM that has received this BS release request changes the buffer set X usage state on the group management table 150 to “unused” (step S2614), and initializes the buffer management table X (step S2615). .

  In addition, the main transmission master CM that has finished the processing of step S2603 changes the buffer set X usage state on the group management table 150 to “unused” (step S2604), and initializes the buffer management table X (step S2604). S2605).

  After that, the main transmission master CM determines whether or not there is a buffer set whose idle reception buffer ID waiting state is “waiting” (step S2606).

  If such a buffer set exists (step S2606; YES), the main transmission master CM changes the waiting state for the empty reception buffer ID related to the buffer set to “not waiting” (step S2607). Then, the main transmission master CM activates an empty reception buffer ID assignment process for the buffer set (step S2608).

  The main transmission master CM that has finished the process of step S2609 determines whether or not the BS free waiting state on the group management table 150 is “waiting for free” (step S2609).

  If the BS idle state is “waiting for free” (step S2609; YES), the main transmission master CM changes the BS idle state to “not waiting” (step S2610). Then, the main transmission master CM activates the BS switching process (step S2611), and then ends this release process.

《Buffer set periodic switching processing》
FIG. 24 is a flowchart of the buffer set regular switching process.

  This buffer set regular switching process is a process (see FIG. 4B) executed by the main reception master CM.

  As shown in FIG. 24, the main reception master CM that has started the buffer set periodic switching process first determines whether or not the negotiation state on the group management table 150 is “negotiated” (step S2001). . Then, when the negotiation state is not “negotiated” (step S2001; NO), the main reception master CM ends this buffer set periodic switching process.

  On the other hand, when the negotiation state is “negotiated” (step S2001; YES), the main reception master CM, based on the BS automatic switching time set in the group management table 150, performs the next execution time of its own processing ( (Date) is set (step S2002). Thereafter, the main reception master CM determines whether or not the BS periodic switching process activation state on the group management table 150 is “activated” (step S2003).

  If the BS periodic switching process activation state is not “activated” (step S2003; NO), the main transmission master CM changes the BS periodic switching process activation state to “activated” (step S2004), and then The buffer set periodic switching process is terminated. On the other hand, if the BS periodic switching process activation state is “activated” (step S2003; YES), the BS switching process for the buffer set whose BS number matches the stored BS number at that time is activated (step S2005). Then, the buffer set regular switching process is terminated.

  As described above in detail, the SS (storage system) 100 according to the present embodiment performs asynchronous remote copy by switching the BS (buffer set) in the same manner as the existing SS 200 described with reference to FIG. It is a system to do. However, the SS 100 is a device that can request the other SS 100 to switch the BS when it is necessary to switch the BS in its own device. It is a switching device.

Therefore, if the SS 100 according to the present embodiment is used, the entire transmission system 100 functions as a single storage system, and write data by the host 300 to a plurality of receiving systems 100 on the remote site side. Thus, it is possible to construct an SS network in which copying with order is performed. Moreover, as is apparent from the functions of the main transmission / reception system 100 described above, the SS 100 is a device (system) capable of remote copying even when only one unit is connected to the host 300. Therefore, according to the SS 100 according to the present embodiment, the storage capacity on the host 300 side can be increased in such a manner that the replacement of the SS 100 already connected to the host 300 is unnecessary.

<Modification>
The SS (storage system) 100 according to the embodiment described above can be variously modified. For example, the SS 100 can be transformed into a device in which the CM 110 that has found that the buffer capacity is insufficient performs the processing of FIG. 17B (and FIG. 17A). Further, the SS 100 can be modified into a device in which the user can set the buffer size.

  Further, the buffer management table (FIG. 8) can be modified such that, for example, a check code of about 8 bytes is stored for each index information. In addition, the size of the buffer management table can be made the same size as the buffer. The size of the buffer management table can be easily adjusted by adjusting the contents and number of index information stored in the buffer management table.

  Further, the SS 100 may be transformed into a device having a specific hardware configuration different from that described above (for example, a device in which only one CM 110 can be mounted), and the configuration / configuration of each table other than the buffer management table Of course, the usage method may be different from the above.

  As described above, the following additional notes are disclosed with respect to the disclosed technology.

(Supplementary note 1) In a storage system connected to a host and a backup storage system,
A first storage unit;
A second storage unit having a plurality of update data storage areas;
A control unit that receives an update data write request to the first storage unit from the host,
The controller is
When an update data write request is received, it is determined whether or not there is a free area having a capacity capable of storing the update data in a processing target storage area that is one of the plurality of update data storage areas. ,
When it is determined that the free area exists, the first write destination information indicating the update data write area in the first storage unit and the update data write in the backup storage system Control information including second write destination information indicating an area and identification information of the update data storage area used as the processing target storage area is stored in the second storage unit, and the update data is stored in the second storage unit. When writing to the first storage unit and determining that the free area does not exist, the processing target storage area is changed to another update data storage area in the second storage unit, and the control information is changed to the first storage unit. 2 storing in the storage unit, writing the update data in the first storage unit, and instructing another storage system connected to the host to change the processing target storage area,
Asynchronously with the processing of the write request, in accordance with the control information stored in the second storage unit, the update data already written in the first storage unit is read and read out in each update data storage area of the second storage unit The storage system, wherein the update data in the update data storage area that has been completed and the second write destination information relating to the update data are transmitted to the backup storage system.

(Supplementary Note 2) The control unit
When the change of the processing target storage area is instructed from the other storage system connected to the host, the processing target storage area is changed to another update data storage area in the second storage unit. The storage system according to appendix 1.

(Supplementary note 3)
Each time a predetermined time elapses, the processing target storage area is changed to another update data storage area in the second storage unit, and the processing target storage area is changed to the other storage system connected to the host. The storage system according to appendix 1 or 2, characterized by:

(Supplementary Note 4) In a storage system connected to a host and a backup storage system,
A first storage unit;
A second storage unit having a plurality of storage area sets each including an index information storage area and an update data storage area;
A control unit that receives an update data write request to the first storage unit from the host,
The controller is
When an update data write request is received, a free area having a capacity capable of storing the update data exists in the update data storage area included in the processing target storage area set that is one of the plurality of storage area sets. Determine whether or not
When it is determined that the free area exists, the processing target storage area in which the first write destination information indicating the write area of the update data in the first storage unit and the update data are to be stored Write request copy control information including storage area designation information indicating a storage area in the update data storage area included in the set is stored in the second storage unit, and the update data is written in the backup storage system When index information indicating an area is stored in the index information storage area included in the processing target storage area set, and the update data is written to the first storage unit, and it is determined that the free area does not exist , Change the processing target storage area set to another storage area set, and store the write request copy control information in the second storage unit The index information is stored in the index information storage area included in the processing target storage area set, the update data is written in the first storage unit, and the processing target is stored in another storage system connected to the host. Instruct to change storage set,
Asynchronously with the processing of the write request, in accordance with each write request copy control information on the second storage unit, update data written in the first storage unit is stored in each update data storage area of the second storage unit. Read the update data on the update data storage area that has been read and the index information on the index information storage area included in the same storage area set as the update data storage area to the backup storage system A storage system characterized by transmitting.

(Supplementary Note 5) In a storage system network including a plurality of host-side storage systems connected to a host and a plurality of backup storage systems, wherein each storage system is connected to be communicable,
The plurality of host-side storage systems are respectively
A first storage unit;
A second storage unit having a plurality of update data storage areas;
A first control unit that receives an update data write request to the first storage unit from the host;
The first controller is
When an update data write request is received, it is determined whether or not there is a free area having a capacity capable of storing the update data in a processing target storage area that is one of the plurality of update data storage areas. ,
When it is determined that the free area exists, the first write destination information indicating the update data write area in the first storage unit and the update data write in the backup storage system Control information including second write destination information indicating an area and identification information of the update data storage area used as the processing target storage area is stored in the second storage unit, and the update data is stored in the second storage unit. When writing to the first storage unit and determining that the free area does not exist, the processing target storage area is changed to another update data storage area in the second storage unit, and the control information is changed to the first storage unit. 2 storing in the storage unit, writing the update data in the first storage unit, and instructing each other host side storage system to change the processing target storage area,
Asynchronously with the processing of the write request, in accordance with the control information stored in the second storage unit, the update data already written in the first storage unit is read and read out in each update data storage area of the second storage unit The update data on the update data storage area that has been completed and the second write destination information related to the update data are transmitted to the backup storage system,
The plurality of backup storage systems are respectively
A third storage unit;
The update data and the second write destination information transmitted by the host side storage system associated with the own backup storage system are received, and the received update data is received by the received second write destination information. A second control unit for writing to a storage area in the third storage unit shown;
A storage system network characterized by comprising:

(Supplementary Note 6) In a storage system network that includes a plurality of host-side storage systems connected to a host and a plurality of backup storage systems, and each storage system is communicatively connected.
The host-side storage systems are respectively
A first storage unit;
A second storage unit having a plurality of storage area sets each including an index information storage area and an update data storage area;
A first control unit that receives a write request for the first storage unit from the host;
The first controller is
When an update data write request is received, a free area having a capacity capable of storing the update data exists in the update data storage area included in the processing target storage area set that is one of the plurality of storage area sets. Determine whether or not
When it is determined that the free area exists, the processing target storage area in which the first write destination information indicating the write area of the update data in the first storage unit and the update data are to be stored Write request copy control information including storage area designation information indicating a storage area in the update data storage area included in the set is stored in the second storage unit, and the update data is written in the backup storage system When index information indicating an area is stored in the index information storage area included in the processing target storage area set, and the update data is written to the first storage unit, and it is determined that the free area does not exist , Change the processing target storage area set to another storage area set, and store the write request copy control information in the second storage unit , Storing the index information in the index information storage area included in the processing target storage area set, writing the update data in the first storage unit, and storing the processing target storage area set in each of the other host side storage systems. Direct the change,
Asynchronously with the processing of the write request, in accordance with each write request copy control information on the second storage unit, update data written in the first storage unit is stored in each update data storage area of the second storage unit. Read the update data on the update data storage area that has been read, and the index information on the index information storage area included in the same storage area set as the update data storage area to the local host-side storage system Send to the associated backup storage system,
The plurality of backup storage systems are respectively
A third storage unit;
The update data and the index information transmitted by the host-side storage system associated with the own backup storage system are received, and the received update data is indicated by the received index information in the third storage A second control unit for writing to a storage area in the unit;
A storage system network characterized by comprising:

100 storage system 105 chassis 110 CM (control module)
111 CPU
112 memory 113 CA
114 RA
115 FC
120 Storage Unit 125 Auxiliary Storage Device 150 Group Management Table 160 Reception Buffer Management Table 180 Initial Copy Session Management Table 190 Write Request Copy Control Table 300 Host

Claims (4)

In the storage system connected to the host and backup storage system,
A first storage unit;
A second storage unit having a plurality of update data storage areas;
A control unit that receives an update data write request to the first storage unit from the host,
The controller is
When the update data write request is received, it is determined whether or not there is a free area capable of storing the update data in the processing target storage area that is one of the plurality of update data storage areas. And
When it is determined that the free area exists, the first write destination information indicating the update data write area in the first storage unit and the update data write in the backup storage system Control information including second write destination information indicating an area and identification information of the update data storage area used as the processing target storage area is stored in the second storage unit, and the update data is stored in the second storage unit. When writing to the first storage unit and determining that the free area does not exist, the processing target storage area is changed to another update data storage area in the second storage unit, and the control information is changed to the first storage unit. 2 storing in the storage unit, writing the update data in the first storage unit, and instructing another storage system connected to the host to change the processing target storage area,
Asynchronously with the processing of the write request, in accordance with the control information stored in the second storage unit, the update data already written in the first storage unit is read and read out in each update data storage area of the second storage unit The storage system, wherein the update data in the update data storage area that has been completed and the second write destination information relating to the update data are transmitted to the backup storage system. The controller is
When the change of the processing target storage area is instructed from the other storage system connected to the host, the processing target storage area is changed to another update data storage area in the second storage unit. The storage system according to claim 1. The controller is
Each time a predetermined time elapses, the processing target storage area is changed to another update data storage area in the second storage unit, and the other storage system connected to the host stores the processing target storage area. The storage system according to claim 1 or 2, wherein a change is instructed. In a storage system network that includes a plurality of host-side storage systems connected to a host and a plurality of backup storage systems, and each storage system is communicably connected.
The plurality of host-side storage systems are respectively
A first storage unit;
A second storage unit having a plurality of update data storage areas;
A first control unit that receives an update data write request to the first storage unit from the host;
The first controller is
When the update data write request is received, it is determined whether or not there is a free area capable of storing the update data in the processing target storage area that is one of the plurality of update data storage areas. And
When it is determined that the free area exists, the first write destination information indicating the update data write area in the first storage unit and the update data write in the backup storage system Control information including second write destination information indicating an area and identification information of the update data storage area used as the processing target storage area is stored in the second storage unit, and the update data is stored in the second storage unit. When writing to the first storage unit and determining that the free area does not exist, the processing target storage area is changed to another update data storage area in the second storage unit, and the control information is changed to the first storage unit. 2 storing in the storage unit, writing the update data in the first storage unit, and instructing each other host side storage system to change the processing target storage area,
Asynchronously with the processing of the write request, in accordance with the control information stored in the second storage unit, the update data already written in the first storage unit is read and read out in each update data storage area of the second storage unit The update data on the update data storage area that has been completed and the second write destination information related to the update data are transmitted to the backup storage system,
The plurality of backup storage systems are respectively
A third storage unit;
The update data and the second write destination information transmitted by the host side storage system associated with the own backup storage system are received, and the received update data is received by the received second write destination information. A second control unit for writing to a storage area in the third storage unit shown;
A storage system network characterized by comprising:

Images